In 2011, 41,146 Americans died by suicide, surpassing the alarming 2010 toll of 38,364. Suicidal behavior (SB) (deaths/attempts) is linked to neurobiological, neurocognitive and behavioral factors, but integrative, multimodal studies are rare. Often, prospective studies, crucial to understanding future risk, have focused on single predictors and a single outcome, implying that SB is homogeneous. Yet, SB is complex and heterogeneous: it can be impulsive or methodically planned; violent or not, reactive to life events or occur despite no obvious stressors suggesting different subtypes. As a first step towards disentangling these subtypes, our model, supported by our pilot data, posits 2 putative phenotypes of SB associated with different patterns of suicidal ideation (SI). We have identified a variable pattern of suicidal ideation typically occurring in individuals with a trauma history wh react to environmental stressors with aggression, pronounced cortisol response and variable SI, in part due to difficulty engaging prefrontal regions in affect regulation. In contrast, we posit tat when SI is elevated but with little fluctuation, it is linked to blunted serotonergic function, and greater cognitive control leading to more planned and lethal SB. The Conte Center (CC) P50-MH090964, conducts cross-sectional studies of SB in Major Depression. We seek to follow this cohort and a newly enrolled sample of depressed patients for 2 years, the highest risk period. Baseline CC assessments mean significant cost savings and maximize use of a rich dataset including: Point-Subtraction Aggression Paradigm for RA, PET imaging with our new agonist, [11C]CUMI-101 to quantify 5-HT1A BPF, and a cognitive emotion regulation task in the MRI scanner. We will administer Stroop and Continuous Performance Tasks to assess cognitive control and prospectively quantify SI and SB, life events and depression using clinical and Ecological Momentary Assessments (EMA). Key innovations are: (1) prospective, multi-modal design to (2) delineate 2 distinct SB subtypes, which may yield more robust predictors; (3) assessment of emotion regulation which may predict SI and SB; (4) use of EMA to delineate origins of SI in real time; (5) Laboratory assessments to characterize trajectories of suicidal subgroups. Impact: In large general population samples, most suicide attempters report SI (NESARC [94.2%]; NLAES [86.8%]), yet >50% of suicides had no prior SB. Thus, prior SB is limited as a predictor of future SB, especially suicide death, so better understanding of SI is crucial to improving prediction and understanding of SB. Critically, if variable SI is a harbinger f impulsive suicide attempts, then delineation of distinct biological and clinical features associate with it or not, may help delineate different risk patterns. Ultimately, it may aid in prospectively identifying those at risk for different expressions/phenotypes of SB. Moreover, future studies could examine the merits of pharmacologic or psychological interventions targeting affect regulation for variable SI, versus antidepressants or cognitive treatments for sustained SI.